Y‐chromosome variability and genetic history of Commons from Northern Italy

Abstract Objectives Genetic drift and admixture are driving forces in human evolution, but their concerted impact to population evolution in historical times and at a micro‐geographic scale is poorly assessed. In this study we test a demographic model encompassing both admixture and drift to the case of social‐cultural isolates such as the so‐called “Commons.” Materials and methods Commons are peculiar institutions of medieval origins whose key feature is the tight relationship between population and territory, mediated by the collective property of shared resources. Here, we analyze the Y‐chromosomal genetic structure of four Commons (for a total of 366 samples) from the Central and Eastern Padana plain in Northern Italy. Results Our results reveal that all these groups exhibit patterns of significant diversity reduction, peripheral/outlier position within the Italian/European genetic space and high frequency of Common‐specific haplogroups. By explicitly testing different drift‐admixture models, we show that a drift‐only model is more probable for Central Padana Commons, while additional admixture (~20%) from external population around the same time of their foundation cannot be excluded for the Eastern ones. Discussion Building on these results, we suggest central Middle Ages as the most probable age of foundation for three of the considered Commons, the remaining one pointing to late antiquity. We conclude that an admixture‐drift model is particularly useful for interpreting the genetic structure and recent demographic history of small‐scale populations in which social‐cultural features play a significant role.

given population, albeit with shifts from one to another depending on the historic/demographic vicissitudes experienced by different human groups. In particular, genetic drift is usually associated with isolated populations, that is groups with a moderate demographic size and low exchange with other populations. However, migration and admixture can play a significant role even in isolates, thus contributing to the peculiar genetic structure that these populations frequently exhibit.
Accordingly, in this study with 'drift' we mean the effects of regional/ local isolation, while 'admixture' more precisely refers to long-range migration events. Well-known examples of this combination are ethnic-linguistic minorities such as the thoroughly investigated Arbereshe and Greek-speaking groups from Southern Italy (Boattini et al., 2011;Destro Bisol et al., 2008;Sarno et al., 2016;Sarno et al., 2017;Sarno et al., 2021;Sineo et al., 2014). The genomic history of these populations may be summarized according to the following scheme: (a) a founding migration event; (b) admixture-to a lesser or higher extent-with local groups; (c) isolation and drift, usually caused by a combination of cultural and geographical factors. Paradoxically, isolation and drift contributed to the conservation, in these groups, of clear genomic traces of their migration history. Furthermore, it has been shown that these populations may function as privileged observatories for detecting and reconstructing specific demographic events, whose genetic traces were diluted or disappeared in more general contexts (Boattini et al., 2015;Gokcumen et al., 2011;Sarno et al., 2016).
Ethnic-linguistic minorities-where the exogenous origin of (at least part of) the group is usually well-documented-are not the only cases to which an admixture-drift model could be applied at a micro-geographic scale. In this study, we investigate the genetic variability of a set of populations which share some aspects with isolates but whose origin is less clear or not known at all. For reasons of simplicity, we will refer to these populations with the term of "Commons," which actually designs a peculiar way of sharing and devolving collective resources. According to the Digital Library of the Commons (http://dlc.dlib.indiana.edu/dlc/), these are juridical institutions based on "shared resources in which each stakeholder has an equal interest." Italy hosts a wide archipelago of Commons, each of them with its own features and history, which makes them a particularly interesting set of populations to explore. Our first research in that vein regarded the Partecipanza of S. Giovanni in Persiceto, a Common located in the Padana Plain whose members form the patrilineal descent of a group of well-identified founder families. Indeed, that study suggested that: (a) an admixture event, probably involving a group of Germanic origins, was at the origin of the Common and (b) subsequent isolation helped to preserve trace of such event in the Y-chromosomal structure of the Partecipanza (Boattini et al., 2015).
Here we focus on four more Commons from the Padana Plain in Northern Italy (Figure 1). Two of them (Nonantola and S. Agata Bolognese) are located in the Central Padana Plain (from now on Central Commons), in the same area of the previously mentioned S. Giovanni in Persiceto, with which they also share the denomination of Partecipanza.
The remaining two (Grignano Polesine and Massenzatica) are at the Easternmost fringes of the Padana Plain, not far from the Po delta (from now on Eastern Commons).
All these Commons share a number of features, most notably all of them evolved from typical medieval contracts called emphyteutic leases, according to which the holder had the perpetual right to the enjoyment of a given property in exchange of some conditions, most notably the increase of the value of the land (ad meliorandum clause), the payment of a small rent and the obligation of stable residence. In our case, emphyteutic concessions were established between important abbeys (Nonantola for Central Commons and Pomposa for Eastern Commons) with vast but deserted properties and local communities as a whole. In exchange of that, the locals contributed significantly to land reclamation and peopling of the territory. At a later time most of these Commons "closed" their ranks and reserved the right to share the collective lands to a restricted group of "founder" families (Arioti et al., 1990;Costato, 1968;Cori, 2011;Fregni et al., 1992). Indeed, three of the four Commons considered here (namely Nonantola, S. Agata B., and Grignano P.) still limit the access to the shared resources to the legitimate, paternal descent of a given set of families/surnames. In addition, and in all the cases, stakeholders must maintain their place of residence within the legal boundaries of the Common. The combined effect of these rules is a long-term relationship between population and land, modeled upon the shared property of the Common.
With this study, and in consideration of the patrilineal nature of most of these Commons, we specifically explore their paternal (Y-chromosome) genetic heritage with the aim of detecting effects of the peculiar socialcultural features related to their presence in the corresponding populations. In particular, we will (a) evaluate if a drift-only model can fully explain their genetic structure or if a model encompassing both admixture and drift provides a better fit and (b) reconstruct their genetic histories and their implication for the wider scenario of Northern Italy.

| The populations
Among the considered Commons (  (Arioti et al., 1990;Fregni, 1992;http://www.partecipanzanonantola.it; https://www.partecipanza.org/). The history of these Commons is quite similar, though the case of Nonantola is best documented. Traditionally, the origins of the Partecipanza of Nonantola are traced back to 1058, when emphyteutic grants were stipulated between the local abbey and the community, but only in 1584 the Common closed its ranks to external families ("chiusura dei ruoli"). This was only the beginning of a bitter fight between two factions, one of them formed by peasants ("bocca viva"), and the other by rich landowners, many of them foreigners ("bocca morta"). Such issue was settled only in 1915, when the descendants of the peasants liquidated the landowners group. In the meantime, in 1856 the Partecipanza separated its administration from that of the municipality and in 1961 finally severed its 1000-year old ties with the Abbey of Nonantola (Venturoli, 2004). S. Agata B. had a less complicated history, being able of redeeming its assets from the Abbey already in 1577 and separating from the municipality in 1797 (Arioti et al., 1990;Fregni, 1992). Both these Commons, as all other Partecipanze, are characterized by a peculiar procedure called cavazione, which is a periodic re-shuffling of shared lands among heads of the participating households. Currently in Nonantola this happens every 18 years and in S. Agata Bolognese every 9 years (Arioti et al., 1990).

The Eastern Commons here considered are Grignano Polesine and
Massenzatica. The Common of Grignano P. is currently named Antichi Beni Originari and it is referred to by locals as Comune. Its shared assets are extended for 130 ha and the number of stakeholders was 327 in 2012 (http://www.antichibenioriginari-grignano.it/). The organization and functioning of the Comune of Grignano P. are very similar to Partecipanze, including a periodic re-shuffling of the shared lands that in this case is performed every 5 years. The Comune originated from an emphyteutic grant from the Abbey of Pomposa probably during the central middle ages. This Common closed its ranks to external families already at the end of 15th century (1494) and redeemed its assets from the Abbey in 1968 (Costato, 1968). As for Massenzatica, the current name of the Common is "Consorzio Uomini di Massenzatica." The origin of this Common is usually associated to an 1182 emphyteutic grant from the abovementioned Abbey of Pomposa. Differently from the Partecipanze (Nonantola, S. Agata B.) and the Comune (Grignano P.), Massenzatica never established a closure of the ranks to foreigners/immigrants neither procedures of re-shuffling among the stakeholders. The right to benefit of the shared goods is here allowed to all male heads of the household permanently residing in the villages of Massenzatica, Monticelli and Italba. The extension of these goods is 353 ha and the number of stakeholders is presently 640. The Common of Massenzatica reached its present form in 1894 with the creation of the "Consorzio" and the resolution of all the easements that, from the times of Pomposa, still weighed on the land (Cori, 2011;http://www.uominidimassenzatica.it/).
From the social-cultural point of view, there were no significant differences between members of Commons and their neighbors: they spoke the same dialect, they shared the same religion and cultural identity. Furthermore, up to 19th century usually Commons were not separated from their municipalities, meaning that local civic administration coincided with that of the Common. Of course, this fact could entail some advantages for Common members, who in addition could count on the availability of unalienable land portions in an eminently agricultural society and on the collaboration/mutual assistance among themselves (Fregni, 1992).
As these advantages disappeared or became less important since the economy of the region shifted towards industry and tertiary, members of Commons however conserved the greatest regard for their ancient institutions and their traditions, which in turn had and still have a great impact in shaping the territory on which the community is settled.
Starting from the late 18th century and especially after the Napoleonic period, Commons experienced long and bitter fights with public authorities, which manifested increasing hostility toward the peculiar way of possessing that these institutions embodied. As a result, some Commons disappeared during this period turning into "normal" private properties. It can be said that the still-existing institutions are only the remnants of a wider phenomenon (Alfani & Rao, 2011;Mantovani, 2017).
Finally, Commons were (and are) not "closed" communities to the "outside world" also from the demographic point of view, as suggested by the relatively low number of endogamic marriages (see also results below), save the partial exception of Nonantola in which an increase of endogamy rates was indeed measured after the "chiusura dei ruoli" in 1584 (Alfani, 2015).

| DNA samples
For each location, we sampled both individuals from the Common and individuals residing in the same place, hence sharing the same All these samples were collected according to the standard "grandparents" criterion (i.e., at least three generations of ancestry in the area of the considered Common) and by excluding related individuals. As for the Control samples, all the selected individuals share surnames that were identified as autochthonous for the considered area by Boattini et al. (2012). Data Whole genome DNA was extracted from buccal swabs by using a salting out protocol modified from Miller (Miller et al., 1988) and quantified with the Qubit ® dsDNA BR Assay Kit (Life Technologies).

| Y-chromosome genotyping
All samples were amplified for the 23 Y-STRs loci included in the PowerPlex ® Y23 System (Promega) following manufacturer recommended protocols. PCR products were sized on an ABI PRISM 310 Genetic Analyzer and alleles were called with GeneMapper ID software (Thermo Fisher Scientific) according to the manufacturer's instructions.
Next, all individuals were additionally genotyped for 42 Y-SNP loci using multiplex SNaPshot mini-sequencing assays (Thermo Fisher Scientific), as described in Sarno et al. (2014). The SNP genotyping was carried out by means of PCR Multiplex amplification, followed by Minisequencing reaction based on dideoxy Single Base Extension (SBE), which was performed with the SNaPshot multiplex kit (Applied Biosystems). SBE products were finally analyzed through capillary electrophoresis on an ABI Prism 310 Genetic Analyser.
Y-STRs data for comparison populations were extracted from the literature for both Italy (Boattini et al., 2013) and Europe (Purps et al., 2014), respectively. In addition, comparison data for within-haplogroup comparisons from 16 Euro-Mediterranean populations (329 samples) were also considered and retrieved from Hallast et al. (2015).

| Pedigree reconstruction
The Partecipanze (Nonantola and S. Agata B.) and the Comune (Grignano P.) kept records of all the households participating to the sharing of leased assets. These records were regularly updated in occasion of the abovementioned periodic re-shuffling of the lands among the householders. Such detailed historic-demographic information allowed the reconstruction of paternal pedigrees up to the late 16th century for Grignano P. and up to 19th century for the Partecipanza of S. Agata B. Unfortunately, records from the Partecipanza of Nonantola could not be accessed after the earthquake of 2012. When more than one individual was found to share a recent paternal ancestor, they were grouped into a single pedigree.

| Generation time
We calculated average generation times based on reconstructed paternal pedigrees, using those individuals for which birth information was available. Rates were calculated by dividing the total number of years for the number of generations encompassed in all the considered pedigrees (after excluding the most remote ancestors, due to uncertainty of their birth date). Confidence intervals (95%) were obtained bootstrapping along different branches of pedigrees (1000 replications).

| Social endogamy
By inspecting all marriages included in reconstructed pedigrees (S. Agata B., Grignano P.), we considered as endogamic those marriages in which both partners bear surnames of the Common.
Marriages for which the surname of the bride was not available were excluded from calculations. Final rates were calculated by dividing the number of socially endogamic marriages for the total number of marriages. Confidence intervals were computed based on a binomial distribution (where the size parameter is equal to the total number of marriages and the probability parameter is given by the observed endogamy rate) using the R function qbinom (R Core Team, 2017).

| Diversity indexes
Standard within-population diversity parameters (Gene Diversity, Mean Number of Pairwise Differences, Nucleotide Diversity) for Y-chromosome haplogroups and STR haplotypes were estimated with Arlequin software 3.5.1.2 (Excoffier et al., 2007). Comparisons of single haplogroup (hg) frequencies between Commons and Controls were performed with Fisher tests and p-values were corrected using the Bonferroni criterion.
The overall differentiation in haplogroup composition between Common-

Control pairs was assessed through AMOVA analysis based on pairwise
Fst statistics as implemented in the above mentioned Arlequin software.

| Multivariate analyses between populations
In order to check the position of our populations within the Italian and European Y-chromosomal genetic landscape, we performed a nonmetric multidimensional scaling (MDS). Because different studies used different levels of hg resolution, the analyses were based on the 23 Y-STR haplotypes that were available for all the considered reference populations. Calculations were performed using the Rst genetic distances computed by the Arlequin software (Excoffier et al., 2007) and the function isoMDS implemented in the R software MASS package (R Core Team, 2017; Venables & Ripley, 2002). The first and the second dimension were represented in a scatterplot, along with the corresponding stress value.

| Within-haplogroup comparisons
In order to explore the genetic variability within haplogroups exhibiting significant frequency differences between Commons and

| Time estimates
Time estimates focused on the haplotype clusters identified by DAPC within the most frequent haplogroups and were limited to clusters including at least nine individuals and separate estimates were also performed for clusters with at least nine individuals within a specific Common. Y-STR mutation rates adopted in the procedure were taken from Ballantyne et al. (2010). Because population events involving Commons are relatively recent, the biasing effect of Y-STRs saturation through time is negligible Boattini et al., 2019) and all Y-STRs (minus DYS385a/b) were therefore used for calculations. In addition, since estimates may be very sensitive to the presence of outliers, we adopted the outlier detection and exclusion procedure described in Boattini et al. (2013). Time estimates were calculated using two different approaches: (a) the SD estimator (Sengupta et al., 2006); (b) the Bayesian method implemented in the Batwing software (Wilson et al., 2003). As for the latter, we adopted a

| Model comparison by approximate Bayesian computation
In order to investigate the recent demographic history of the considered groups, we compared three different models for each Common/ Control pair. All models follow the same demographic history these percentages are 50% and 5%, respectively (Figure 2b). For a detailed description of all models' parameters see Table S1.
We generated 5000 simulated data sets ( In order to better define the patterns emerged from models' comparison, we performed the estimation of the admixture proportion in the considered Commons (introgression from Pop 2 to the Common).
To do this, we designed a demographic model in which the admixture parameter is free to vary at each iteration from 0% to 50%. We performed 500,000 simulations and then estimated the posterior distribution through the ABCestimator tool of the ABCToolbox package (Wegmann et al., 2009).

| Diversity indexes
We calculated classic indexes of genetic diversity for all the considered populations (Commons and Controls) using both Y-STR haplotypes and haplogroup frequencies (

| DAPC and time estimates
Within-haplogroup analyses were performed for the most frequent haplogroups in the considered Commons, that is, R-U152*, R-L51, R-L2, T-M70 and G2a-U8 (Table S5, Figure S2). DAPC results ( Table 2,   Table S5) show that these haplogroups revealed a number of welldefined clusters variable from 3 to 7. In addition, all of them exhibit at least one Common-specific cluster, that is, a cluster in which the majority (>50%) of the haplotypes is associated to a given Common.
More precisely, S. Agata B. presents two associated clusters, both of them within R-U152* (namely 4 and 5); Nonantola also presents two These results suggest that each of these Common-specific clusters may have spread from a common recent ancestor that could have been living around or after the time of separation between Commons and Controls. Accordingly, we estimated the ages of all the identified clusters whose size exceeds nine individuals, both using Common-specific individuals only (when available and in sufficient number) and the whole clusters. Being aware that time estimates-and particularly STR-based estimates-must be taken with caution, we interpret Common-specific estimates as lower limits to the age of foundation/segregation of Commons, while whole-cluster estimates provide an idea of the temporal depth of the "general" genetic background.
In general, our results (

| Model comparison
We then compared three demographic models for interpreting the observed differences between Commons and Controls ( Figure 2).
According to the first model, such differences are explained solely by drift after the establishment of the Commons (Figure 2a). The second and the third models instead add an external contribute to the genetic make-up of the Commons, considering respectively a 30% and 50% admixture component (Figure 2b). Results for each of the Common/ Control pair are summarized in Table 3, where each panel reports the classification error obtained from the comparison of the set of models.
Firstly, we compared the three considered demographic models in a single run (Table 3, upper panel), which resulted in high classification errors, possibly due to high similarity between 30% and 50% admixture models. Accordingly, we decided to separately compare the Drift-only model with each of the two "Admix" models (30% and 50%; Table 3 the considered models proved to be able to reproduce the observed variation, as shown by linear discriminant analysis (LDA; Figure S3).

| Parameters estimation
When performing model selection, we tested demographies accounting for different proportions of admixture (0%, 30%, and 50%) from an external source of genetic variation. This was done in order to maximize the identifiability of the considered models, which correlates with the degree of differentiation among the tested demographic histories. However, the admixture parameter could actually vary in a wider range than that explored by the above tested models. Accordingly, we estimated admixture rate for all the considered Commons using ABC (Table 4, Figure S4). Consistently with the model selection analysis, posterior probabilities in Nonantola and S. Agata B. appear narrower and shifted toward lower values than those observed for Grignano P. and Massenzatica. Median admixture rates in the Eastern  Commons (Grignano P. and Massenzatica) were around 30% (with mean value of 56%) in Grignano P. and 20% (with mean value of 41%) in Massenzatica. In Central ones (S. Agata B. and Nonantola) they were around 10% with slightly narrower confidence intervals. In other words, a small amount of admixture/introgression in Nonantola and S. Agata B., which showed support for the model with "only drift," is still compatible with the observed data.

| DISCUSSION
The main aim of this study is to understand how admixture and drift may contribute to the genetic make-up of peculiar populations such as those associated to the presence of Commons. At the same time, these groups, investigated at a micro-geographic scale after a careful sampling, may reveal aspects of the recent genetic history of the T A B L E 4 ABC estimator results describing admixture estimates between a non-Italian ghost populations (Pop 2; see Figure 2) and each Common, showing mean, mode, median, and 95% highest posterior density (HPD) interval region in which they are located-aspects that are normally hidden or more difficult to detect in the "general" population. Similar considerations were also expressed in a study about four Anatolian villages by Gokcumen et al. (2011), according to whom "broad, ethnicity-based sampling is inadequate to capture the genetic signatures of recent social and historical dynamics, which have had a profound influence on contemporary genetic and cultural regional diversity".
Indeed, a previous study about the Partecipanza of S. Giovanni in Persiceto (Boattini et al., 2015) suggested that social-cultural features such as the presence and the persistence through the centuries of Commons seem particularly apt to produce, as a biological effect, peculiar micro-geographic genetic structures in which both admixture and drift could have played an important role. Accordingly, we sampled four Commons from the Padana Plain, Northern Italy, each of them characterized by its peculiar set of rules to keep the property of the shared goods in the hands of a restricted and well-defined group of stakeholders ( Figure 1, Table 1). Then, we compared their Ychromosomal structure to that of Control populations, that is, groups that share the same environmental and cultural features of the Common, excepted for the affiliation to the Common itself.
Our results showed that in all cases Commons are characterized by a significant reduction of genetic diversity, compared with their Controls (Figure 3, Table S3). This was somehow anticipated, since the social-cultural rules determining the affiliation to a given Common are likely to have caused a certain degree of isolation between the Common and the neighboring populations, at least from the paternal side. Such rules are particularly stringent for Nonantola, S. Agata B. and Grignano P., where both local residence and legitimate patrilineal descent are required, while in Massenzatica only the former is applied (Table 1). Accordingly, only a moderate diversity reduction was observed in Massenzatica, while such phenomenon is more evident in the three remaining Commons, in addition showing a clear association with population size. Furthermore, as revealed by AMOVA, Control groups show an overall higher similarity between them than Commons do. A significant reduction of Y-chromosomal diversity was not detected in the Partecipanza of S. Giovanni in Persiceto, but this fact could be due to the lower number of Y-STRs and Y-SNPs typed in that study (Boattini et al., 2015).
Differences between Commons and their Controls are also apparent when inspecting their haplogroup composition (Table S4). In fact, in all cases we detected an overall significant difference in terms of Fst between the considered pairs of populations. In particular, few haplogroups clearly characterize each of the considered populations.
Indeed, no significant difference was observed in S. Giovanni in Persiceto between the Common and its Control from the mtDNA point of view (Boattini et al., 2015).
We then explicitly tested the admixture-drift model in our populations via ABC. A similar approach was recently used at a larger geographic scale in Kutanan et al. (2019), which however is here aimed at disentangling specific micro-geographic patterns and reconstructing recent demographic histories. In particular, we considered three different models namely: (a) Drift-only, (b) Admix-30 (i.e., 30% admixture followed by drift), and (c) Admix-50 (i.e., 50% admixture followed by drift) (Figure 2). Our simulations suggested that the Drift-only model is favored for Central Commons (Nonantola, S. Agata B), while for Eastern Commons (Grignano P., Massenzatica) some degree of introgression from an external source of genetic variation is more probable (Table 3, Figure S3). However, admixture cannot be excluded also for Central Commons, albeit with a much lower admixture rate, as suggested by ABC parameter estimation ( Figure S4). In fact, after estimating the admixture parameter, we obtained values around 17%-21% with large confidence intervals for Eastern Commons, and 11%-15% with tight confidence intervals for Central ones (Table 4) Our second aim was to use this information in order to reconstruct the genetic history of these groups in the wider context of Italian genetic history. Accordingly, we proceeded to a more detailed analysis of the detected haplotypes/haplogroups/paragroups using DAPC (Table 2, Table S5, Figure S2). Our results revealed 11 Common-specific clusters of haplotypes within the most frequent haplogroups. The great majority (9/11) of them were exclusive or almost exclusive of one of the considered Commons, albeit with two exceptions, namely Cluster 3 in paragroup R-U152*, which is associated to the Nonantola Common but is sporadically observed in other Commons and Controls, and Cluster 2 in hg R-L51*, which is associated to the Common of Massenzatica, but is sporadically observed in all Controls and also in three central-northern European populations (Bavarian, Danish, English). Interestingly, Massenzatica and Nonantola are also the Commons that showed the lowest amount of diversity reduction, suggesting a lower degree of paternal isolation than the other ones ( Figure 3, Table S3).
In general, it seems plausible that Common-specific clusters would mark expansion events within the corresponding populations, therefore estimates of their time depth should work as a lower bound for the time of origin of the Common itself.
A Despite considering the fact that STR-based estimates should be taken with the greatest caution, our SD and Batwing time estimates, besides being in agreement with each other, seem to be coherent with historical information about the origin of these communities (Arioti et al., 1990;Cori, 2011;Costato, 1968;Fregni, 1992;Venturoli, 2004). Indeed, dates (Table 2) range between central Middle Ages and early Modern Age, which is in agreement with an origin of the Commons at least 1000 years ago and following re-expansion events after the segregation of the founder families (in Nonantola, S. Agata B., Grignano P.) around 500 years ago. Again, Massenzatica is an exception pointing towards an earlier origin-likely in Late Antiquity-which seems coherent with local archaeological remains (Cori, 2011). Whole clusters age estimates instead point especially to the interval between 2000 and 5000 years before present, which is in agreement with the more general make-up of the Italian population (Antonio, 2019;Boattini et al., 2013;Cocca, 2020;Fernandes et al., 2020;Marcus et al., 2020;Sarno et al., 2017;Sazzini et al., 2016;Sazzini et al., 2020).
Recent studies based on modern and ancient genomes suggested that the "core" of Italian genetic variation was already in place in the early antiquity. Late antiquity and medieval migrations, as far as we are concerned, seem to have left only minor traces in the Italian genomic background, which are not apparent in "general" populations (Antonio, 2019). For instance, traces of late medieval migrations from Southern Balkans to Southern Italy are detectable in some ethniclinguistic minorities that still conserve their original language (Arbereshe) but not in the "average" Southern Italian population Sarno et al., 2017). Similarly, the Partecipanza of S. Giovanni in Persiceto was hypothesized to spot traces of an early medieval migration from Northern Europe which were lost in other groups (Boattini et al., 2015).
As above discussed, our analyses suggested that Eastern Commons, that is, Grignano P. and Massenzatica, are the most likely cases for admixture-drift, while the Y-chromosomal variability of Central Commons (Nonantola and S. Agata B.) is more easily explained by drift only. In light of these results together with DAPC and dating experiments, the most likely scenario could be the following. The three Commons with the patrilineal descent rule (Nonantola, S. Agata B., and Grignano P.) were founded in the central Middle Ages, while Massenzatica could have been a few centuries older. All communities likely stem from local populations that lived in the area from a long time, however Grignano P. and Massenzatica probably incorporated a~20% contribute from an external population around the same time in which the Common was founded. Unfortunately, it is not possible to be more precise about the origin/identity of these external populations, limiting ourselves to the observation that Grignano P. exhibits an otherwise rare haplogroup, T-M70, which could refer to a Mediterranean background, while R-L51*, which is typical of Massenzatica, is mostly observed in central Europe (Busby et al., 2012;Harney et al., 2018;Mendez et al., 2011;Myres et al., 2011).
Later, isolation and genetic drift were induced in Nonantola, S. Agata B., and Grignano P. in virtue of their patrilineal separation from the neighboring populations. In Massenzatica, instead, we believe that the peculiar environment in which the community was founded-a lagoon-like environment subjected to frequent floods from the Po river-played a fundamental role in its segregation.

| CONCLUSION
This study shows how the co-presence of admixture and drift forms a suitable model for explaining the genetic variability of at least two of the four considered Commons, namely Grignano P. and Massenzatica.
At the same time, we observed that the peculiar social-cultural features of Commons-based on patrilineal descent and local residenceinfluence their Y-chromosomal variability in a way reminiscent of ethnic-linguistic minorities, where phenomena such as isolation and/or admixture are frequently observed. The collected results allowed to reconstruct some aspects of the genetic history of the con- Conceptualization; data curation; formal analysis; investigation; methodology; supervision; writing-original draft.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are openly available in YHRD at yhrd.org, reference numbers YA004724 and YA004725.